Efficient Feature Engineering and Classification for Large-Scale Email Spam Detection

Authors

  • Moavia Hassan University of Engineering and Technology, Taxila, Pakistan.
  • Muhammad Ilyas University of Engineering and Technology, Taxila, Pakistan.
  • Qamas Gul Khan Safi University of Engineering and Technology, Taxila, Pakistan.
  • Muhammad Munwar Iqbal University of Engineering and Technology, Taxila, Pakistan.
  • Farooq Ali University of Engineering and Technology, Taxila, Pakistan.

DOI:

https://doi.org/10.54692/lgurjcsit.v10i1.191

Keywords:

'Spam Detection', 'Machine Learning', 'Logistic Regression,', 'SpamBase Dataset,', 'Random Forest'

Abstract

Email spam comprises 45% of global email traffic. It poses significant challenges, including resource wastage and security risks, such as Phishing and Malware Distribution. The implementation of modern machine learning algorithms has demonstrated their effectiveness in identifying and filtering spam email messages. Existing research has employed various techniques for classifying spam email datasets. The Proposed framework addresses spam detection at a large scale using supervised machine learning techniques. Exploratory Data Analysis was performed to understand data characteristics and identify patterns relevant to spam classification. Data preprocessing techniques were applied to handle missing values, normalize features, and remove noise. Feature engineering was then carried out to generate meaningful attributes that improve model performance. Then, machine learning algorithms were trained and tested for classification evaluation. The Random Forest classifier achieved the best performance, with an accuracy of 97.86% and an ROC-AUC score of 0.97. The findings highlight the importance of feature engineering and Random Forest methods in spam detection tasks. This work provides practical insights for building accurate and scalable spam filters. It also opens up future research directions involving advanced feature engineering and deep learning approaches to further enhance classification performance.

Downloads

Published

08-06-2026

Issue

Section

Articles